Potential conflict of interest: Dr. Gilroy is on the speakers' bureau of Genentech, Gilead, and Vertex.
Supported by National Institutes of Health grants CA 53596, DK092100, and P20RR021940.
The current study tests a hypothesis that nuclear receptor signaling is altered in chronic hepatitis C patients and that the altered pattern is specific to alcohol drinking history. The expression of a panel of more than 100 genes encoding nuclear receptors, coregulators, and their direct/indirect targets was studied in human livers. Gene expression pattern was compared between 15 normal donor livers and 23 hepatitis C virus (HCV) genotype 1–positive livers from patients without a drinking history (matched for age, sex, and body mass index). HCV infection increased the expression of nuclear receptors small heterodimer partner and constitutive androstane receptor (CAR) as well as genes involved in fatty acid trafficking, bile acid synthesis and uptake, and inflammatory response. However, the expression of retinoid X receptor (RXR) α, peroxisomal proliferator-activated receptor (PPAR) α and β as well as steroid regulatory element-binding protein (SREBP)-1c was decreased in HCV-infected livers. Gene expression pattern was compared in chronic hepatitis C patients with and without a drinking history. Alcohol drinking increased the expression of genes involved in fatty acid uptake, trafficking, and oxidation, but decreased the expression of genes responsible for gluconeogenesis. These changes were consistent with reduced fasting plasma glucose levels and altered expression of upstream regulators that include RXRα, PPARα, and CAR. The messenger RNA levels of fibroblast growth factor 21, interleukin-10, and fatty acid synthase, which are all regulated by nuclear receptors, showed independent correlation with hepatic HCV RNA levels. Conclusion: Our findings suggest that those genes and pathways that showed altered expression could potentially be therapeutic targets for HCV infection and/or alcohol drinking-induced liver injury. (HEPATOLOGY 2011)
Hepatitis C is the principal cause of death from liver disease and the leading indication for liver transplantation in the United States.1 Advances have been made in antiviral treatment with the combination of pegylated interferon and ribavirin, but less than half of the patients infected with genotype 1 achieve sustained virological response (SVR).2–4 With the recent development of hepatitis C virus (HCV) protease inhibitors (telaprevir and boceprevir), only about 70% of the treatment-naïve patients and half of the patients who failed standard treatment achieve SVR.5–8 There is an urgent need to understand the virus-host interaction in order to develop novel intervention strategies.
An intriguing feature of HCV infection is its relationship with lipids, as indicated by the following: (1) HCV virions circulate in serum bound to lipoproteins, called lipoviroparticles;9 (2) steatosis is prevalent in HCV-infected patients;10, 11 and (3) lipids are essential for the HCV life cycle and the virus was named a “metabolovirus.”12, 13 Nuclear receptors, which are transcriptional factors, play pivotal roles in lipid homeostasis. In addition, nuclear receptors also play important roles in regulating inflammatory response and fibrogenesis.13–15 HCV infection is associated with changes in nuclear receptor-mediated signaling. However, because various in vitro and animal models were used for most of the studies, inconsistent findings were obtained.13, 15, 16 The goal of the current study was to use human livers to test a hypothesis that HCV infection is associated with alteration of hepatic nuclear receptor-mediated pathways, which may in turn contribute to viral replication and the pathological process.
At least moderate alcohol consumption is found in two-thirds of patients with chronic hepatitis C, and only half of them stop alcohol drinking upon counseling and initiation of hepatitis C treatment (www.easl.eu/_clinical-practice-guideline). Heavy alcohol intake is associated with an accelerated fibrosis progression, a higher incidence of cirrhosis and hepatocellular carcinoma (HCC), and a lower rate of SVR.17, 18 Nuclear receptor-mediated pathways not only play a role in alcohol detoxification, but also contribute to alcohol-induced liver pathogenesis in animal models.19, 20 Thus, another goal of this study is to identify biomarkers for alcohol drinking in HCV-infected patients.
We identified differential expression genes encoding hepatic nuclear receptors, coregulators, and their downstream targets in hepatitis C patients and normal liver donors. Differential gene expression was also examined in chronic hepatitis C patients with and without a history of alcohol drinking. Our data showed that the expression of many studied genes correlated with hepatic HCV RNA level. Our findings suggest that nuclear receptor-mediated pathways play an important role in HCV replication and pathogenesis and thus can be potential therapeutic targets to control the disease process.
Forty-four liver specimens were obtained from the University of Kansas (KU) Liver Center Tissue Bank (http://www.kumc.edu/livercenter/liver_tissue_bank.html). Consent was obtained from all patients according to a protocol approved by the Institutional Review Board. These specimens were from patients with genotype 1 HCV infection. Inclusion criteria were as follows: patients older than 18 years and positive for both anti-HCV antibody (Abbott ARCHITECT anti-HCV test) and serum HCV RNA (Roche Cobas Ampliprep). The following patients were excluded: those positive for hepatitis B virus surface antigen; those with primary biliary cirrhosis, autoimmune hepatitis, Wilson's disease, hemochromatosis, or coinfection with human immunodeficiency virus; or those who had been treated with antiviral or immunosuppressive agents within 6 months of when liver tissues were obtained. Among the 44 HCV-infected patients, 23 patients were not current or former alcohol users (group A). Thirteen individuals in group A were male (group A1). Twenty-one of 44 HCV-infected patients were either current or former alcohol drinkers and they were all male (group B). Group A1 and B were matched for sex, age, and body mass index (BMI) (Table 1). Fifteen liver specimens obtained from the donors at the Liver Transplant Program at the University of Kansas Hospital were used as normal controls (group C). Groups A and C were matched for age, sex, and BMI.
Table 1. Patient Characteristics
C (Normal livers from donors)
A (HCV patients without a reported history of drinking)
A1 (male) (HCV patients without a reported history of drinking)
B (male) (HCV patients with a reported history of drinking)
P < 0.05, comparison between group A1 and group B.
The data for other cases are not available.
‡Patients with cirrhosis are not required to report histological activity.
Serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin (TBILI), alkaline phosphatase (ALP), total cholesterol (CHOL), triglyceride (TRIG), and fasting plasma glucose were obtained from patients' charts and all the tests were performed within 3 months of liver biopsy. The HCV genotype was determined by sequencing using the TRUGENE HCV 5′NC Genotyping Kit.
Hematoxylin and eosin–stained as well as Masson's trichrome-stained liver sections were used for diagnosis by the pathologists. The degrees of inflammation and fibrosis were evaluated according to the criteria proposed by Ishak et al.21 Steatosis was graded based on percentage of hepatocytes involved: none (<5%), mild (5%-33%), moderate (≥33%-66%), or severe (≥66%).
Hepatic Messenger RNA Quantification.
Hepatic RNA was extracted for study of gene expression by real-time polymerase chain reaction (PCR). The studied genes are listed in Supporting Table 1. Data were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) messenger RNA (mRNA) level.
Student t test was used for gene expression comparisons between two groups. For the correlation analysis, comparative cycle threshold (Δ Ct) values were used. Pearson correlation analysis was used to study the correlation between gene expression and hepatic HCV RNA. (According to the Kolmogorov–Smirnov Z test, the Δ Ct data are within normal distribution.) Multivariate linear regression analysis was used to identify the independent correlations for genes that had significant correlation as identified by bivariate correlation analysis. P < 0.05 was considered statistically significant.
Characteristics of Patients Studied.
Demographic information and clinical data of the 44 studied patients and 15 liver donors are summarized in Table 1. Except for fasting plasma glucose level, which was reduced in patients with a drinking history, other parameters were not different between Group A1 and B or between Group C and A. Most patients in Group B had a heavy drinking history and were binge drinkers. Only 28.6% patients reported that they were current drinkers (Table 2).
Table 2. Alcohol Consumption Profile in Group B (n = 21 Men with HCV Infection)
Heavy drinking: three or more drinks on most days or every day in a week; moderate drinking: two drinks on most days or every day in a week; light drinking: no more than one drink on most days or every day in a week.
Consumption of five or more drinks during a single occasion for men or four or more drinks during a single occasion for women.,
Years of alcohol consumption
Range: 3-46 (28.00 ± 11.64)
Drinking onset age, years
17.57 ± 3.57
6/15 (28.6% current)
Average drinks per day (light/moderate/heavy drinking)*
Among the 17 nuclear receptors and 5 coregulators studied (Supporting Table 1), the nuclear receptors retinoid X receptor (RXR)α; peroxisome proliferator-activated receptor (PPAR)α and β, small heterodimer partner (SHP), and constitutive androstane receptor (CAR) as well as the coregulator nuclear receptor coactivator 3 (NCOA3) showed a difference at the mRNA levels (Fig. 1A). SHP, CAR, and NCOA3 mRNA levels were increased in HCV-infected patients in comparison with normal controls. Hepatic RXRα, PPARα, and PPARβ mRNA levels were decreased in HCV-infected patients compared with normal controls. There was no change in the expression levels of other nuclear receptors including FXR (Supporting Fig. 1A).
The expression of genes (Supporting Table 1) that play pivotal roles in regulating lipid homeostasis was studied. The mRNA level of steroid regulatory element-binding protein (SREBP)-1c was decreased in HCV-infected livers (Fig. 1B). The expression levels of SREBP-1c target genes (fatty acid synthase [FAS], acyl-CoA carboxylase [ACC], and fatty acyl-CoA elongase [FAE]) were modestly reduced in HCV-infected livers, but did not reach statistical significance (Supporting Fig. 1B). Genes responsible for fatty acid uptake (fatty acid transport protein 2 and 5 [FATP2 and FATP5]) and glucose uptake (facilitated glucose transporter 2 [GLUT2]) were up-regulated in HCV-infected livers (Fig. 1B). The expression of low-density lipoprotein receptor (LDLR), which plays a key role in viral entry to the hepatocyte,22 was decreased by four-fold in HCV-infected patients (Fig. 1B). In addition, the mRNA levels of microsomal triglyceride transfer protein (MTP), which is essential for very low-density lipoprotein (VLDL) and HCV secretion from the infected cells,23 was increased in HCV-infected livers (Fig. 1B).
Cytochrome P450, family 7, subfamily A, polypeptide 1 (CYP7A1) catalyzes a rate-limiting step in cholesterol catabolism and bile acid biosynthesis. Na+/taurocholate cotransporter (NTCP) controls the uptake of bile acid. The expression levels of both genes were increased in HCV-infected livers (Fig. 1C). The mRNA level of CYP7A1 increased more than 13-fold. CYP7A1 and NTCP are negatively regulated by SHP. However, HCV-infected patients have increased CYP7A1 and NTCP as well as SHP mRNA levels.
Inflammatory pathway gene expression was altered in HCV-infected livers (Fig. 1D). HCV-infected patients had increased expression of tumor necrosis factor α (TNF; three-fold) and nitric oxide synthase 2 (NOS2; seven-fold). The expression of genes in the fatty acid oxidation pathway and antioxidant system showed no significant difference (Supporting Fig. 1C,D).
Biomarkers for Alcohol Drinking in HCV-Infected Men.
There is a sex difference in alcohol intake and alcoholic liver disease; female sex is a protective factor for drinking and yet female individuals are more susceptible than male individuals to the development of alcoholic liver disease.19 It is essential to use sex-matched populations to study biomarkers for alcohol use. Thus, we compared gene expression patterns between HCV-infected men with and without an alcohol drinking history. In patients who had a history of drinking, hepatic RXRα and PPARα mRNA levels were increased whereas CAR mRNA level was decreased in comparison with those who did not have such a history (Fig. 2A). RXRα mRNA levels increased more than 2.5-fold, implying the importance of retinoid signaling as a response to alcohol drinking. In addition, liver X receptor (LXR), retinoic acid receptor (RAR)α, and nuclear receptor subfamily 1, group D, member 2 (Rev-Erb)β mRNA levels were different between these two cohorts (Fig. 2A). LXR plays a key role in fatty acid synthesis and regulates the expression of SREBP-1c.24, 25 Rev-Erbβ negatively regulates the expression of CD36, fatty acid binding protein 3 and 4 (FABP3 and FABP4), uncoupling protein 3, SREBP-1c, and stearyl-CoA dehydrogenase (SCD-1).26 The decreased Rev-Erbβ is consistent with the up-regulation of CD36 and FABP3 (Fig. 2C). NCOR2 and NCOA3 mRNA levels were significantly different between the two groups. Patients who had a drinking history had decreased NCOR2 and NCOA3 mRNA levels (Supporting Fig. 2A).
Consistent with the changes in RXRα and PPARα, the expression levels of genes related to fatty acid oxidation were increased in patients with alcoholism (Fig. 2B). These up-regulated genes are involved in the mitochondrial β oxidation pathway (hydroxyacyl-CoA dehydrogenase [HADH]α and acyl-CoA dehydrogenase [ACADS]), peroxisomal oxidation pathway (acyl-CoA oxidase 1 and 2 [ACOX1 and 2]), and microsomal oxidation pathway (CYP2E1 and CYP4A11). Intriguingly, gene expression in the antioxidant and inflammatory systems did not change significantly (Supporting Fig. 2B).
In the fatty acid uptake and intracellular trafficking pathway, CD36 and FABP3 mRNA levels were increased in patients who had a history of drinking (Fig. 2C). There was no change in the expression of genes that are involved in the fatty acid synthesis or VLDL secretion pathways (Supporting Fig. 2C-E). In the hepatic gluconeogenesis pathway, both glucose-6-phosphatase (G6P) and phosphoenolpyruvate carboxykinase (PEPCK) mRNA levels were reduced in alcoholic patients (Fig. 2D). These changes along with the reduction of GLUT2 mRNA level are consistent with the reduced plasma glucose level found in alcoholic patients (Supporting Fig. 3).
Correlations Between Hepatic HCV RNA Level and the Expression of Hepatic Genes.
Using bivariate correlation analysis, the mRNA levels of PPARγ, RARβ, RARγ, liver receptor homolog-1 (LRH-1), farnesoid X receptor (FXR), SCD1, FAS, fibroblast growth factor 21 (FGF21), G6P, IL-10, and retinoid-inducible gene 1 protein (RIG1) were correlated with hepatic HCV RNA levels. All the correlation coefficients were higher than 0.4, and RARγ had the best correlation coefficient (0.57) (Table 3). Stepwise multivariate linear regression analysis showed that FGF21, IL-10, and FAS mRNA levels were independently correlated with hepatic HCV RNA (Table 4). The adjusted R2 of this model was 0.63. Predictability is shown in Fig. 3.
Table 3. Bivariate Correlation Analysis for HCV RNA and Gene Expression
P value (Two-Tailed)
Lipid and glucose metabolism pathway
Immune response and inflammatory pathway
Table 4. Multivariate Linear Regression Analysis for HCV RNA
The molecular mechanisms involved in HCV disease progression are not well understood. Our data indicate that viral infection can alter nuclear receptor-mediated signaling, which may lead to HCV infection-associated pathological conditions. Nuclear receptors are popular drug targets, and drugs that modulate nuclear receptor activity are among the most prescribed pharmaceuticals on the market.27 Nuclear receptors may have the potential to be therapeutic targets to control the HCV-associated disease process.15, 16
Among the nuclear receptor genes studied, RXRα, PPARα and β, SHP, and CAR showed significant changes in their expression levels in HCV-infected livers. These nuclear receptors play important roles in lipid metabolism or related pathways (Fig. 4). Our published data show that serum cholesterol and triglyceride levels are increased due to hepatocyte RXRα deficiency.28, 29 PPARα modulates genes encoding lipid metabolism enzymes, lipid transporters, and apolipoproteins, and PPARα mRNA level is reduced in HCV-infected livers. An interesting finding is the concomitant down-regulation of RXRα and PPARβ in HCV-infected patients. It has been shown that the expression of RXRα is not altered in HCV-associated cirrhotic livers.30 The difference may be due to lack of advanced cirrhosis in most patients included in the current study. PPARβ has recently received a great deal of attention in research on metabolic diseases. Activation of mouse PPARβ increases fatty acid β-oxidation, which is accompanied by marked reduction of lipid droplets in skeletal muscle.31 Hepatic-restricted PPARβ activation produces hepatic glycogen and an increase in monounsaturated fatty acid production as well as up-regulation of glucose utilization.32
Our results showed that hepatic PPARβ mRNA was decreased and GLUT2 mRNA was increased in hepatitis C patients, suggesting an imbalance in using fat and sugar as energy sources in HCV-infected livers. Coregulator NCOA3 mRNA level was increased in HCV-infected livers, which is consistent with the finding that deletion of NCOA3 in mice prevents high-fat-diet-induced steatosis and inflammation.33 However, whether the modest changes in NCOA3 mRNA level found in patients are biologically significant remains to be validated.
In vitro, HCV core proteins NS2 and NS5 induce hepatic lipid accumulation by activating SREBP-1c and PPARγ.34–36 However, SREBP-1c mRNA level was reduced in HCV-infected livers. Our unpublished data also show reduced expression of SREBP-1c in HCV core protein transgenic mice. The expression of SREBP-1c is negatively regulated by SHP and CAR and positively regulated by PPARβ (Fig. 4). The up-regulated SHP and CAR as well as the down-regulated PPARβ found in our HCV-infected patients could explain reduced SREBP-1c expression. Reduced expression of SREBP-1c could also be due to overloading of fat in the hepatocyte caused by up-regulated expression of FATP2 and FATP5, and thus potentially is an adaptive response.
SHP plays a central role in lipid and glucose metabolism by inhibiting SREBP-1c, MTP, NTCP, and CYP7A1 expression and increasing G6P and PEPCK expression.13, 37 In HCV-infected patients, increased hepatic SHP, MTP, NTCP, and CYP7A1 mRNA was observed, and FXR, G6P, and PEPCK mRNA levels did not change. This finding suggests that the FXR-SHP-CYP7A1 regulatory loop is totally compromised in HCV-infected liver. The observed changes could be due to HCV infection. Alternatively, such changes could be adaptive host responses in order to minimize liver injury. MTP is essential for hepatic lipoprotein assembly and secretion, and VLDL is important for HCV secretion from the infected cells.23 In addition, bile acid via FXR promotes genotype 1 HCV replication.38, 39 Thus, all these alterations are related to HCV life cycle.
Activation of CAR ameliorates hyperglycemia by suppressing glucose production and stimulating glucose uptake and usage in the liver and improves steatosis by inhibiting hepatic lipogenesis and inducing β-oxidation.40 In our hepatitis C patients, CAR was significantly up-regulated and this was accompanied by decreased SREBP-1c and increased GLUT2 expression. This finding suggests that CAR may play a significant role in lipid and glucose metabolism in HCV-infected livers.
In ethanol-fed mice, hepatic PPARα-mediated signaling is decreased.41–43 In addition, AMPK activity and fatty acid synthesis-related genes are down-regulated.44 In the HCV-infected patients who had a history of drinking, our results showed that PPARα and RXRα expression levels were increased, with concomitant up-regulation of their target genes involved in fatty acid oxidation and hepatic uptake and intracellular trafficking. Species difference may account for the differential findings. There were both current and noncurrent drinkers in group B, but no significant difference could be found in gene expression between the two groups (Supporting Table 2). This suggests the possibility of active drinking in “noncurrent drinkers”. In addition, the gene expression alteration does not seem to be caused by differences in disease severity because there was no difference in liver panel, severity of fibrosis, or inflammation in these two cohorts (Supporting Table 3). Although PPARα and RXRα and their target genes were up-regulated in patients with a history of alcohol drinking, the genes involved in antioxidant and inflammatory pathways did not change their expression level significantly (Supporting Fig. 2B). This result does not support the hypothesis that alcohol and HCV synergize through increasing PPARα activity, lipid peroxidation, oxidative stress, and thus liver injury. Other mechanisms have been proposed to explain the synergism of HCV infection and alcohol intake. For example, alcohol impairs the intracellular innate immune response in human hepatocytes and promotes HCV infection and replication.45
Multivariate analysis showed an independent association between the hepatic mRNA levels of FAS, FGF21, and IL-10 with HCV RNA. All these genes are regulated by nuclear receptors or coregulators. FAS expression is regulated by SREBP-1c, which is regulated by LXR and PPARγ.13, 24, 25 FGF21 is involved in fat oxidation or lipolysis in the liver and is a target of PPARα.46–48 IL-10 is an anti-inflammatory cytokine and down-regulates Th1 effector mechanisms. The expression of IL-10 in hepatocytes is increased by fatty acids and such regulation is mediated by PPAR-γ coactivator 1α (PGC-1α).49 The relationship between FAS and HCV replication has been studied in a subgenomic replicon system50 and in the JFH1 infectious system.51 Inhibition of fatty acid synthase (FAS) by cerulenin or C75 blocks HCV replication. There have been no reports on the relationship between FGF21 and HCV replication. Our result for the first time showed a negative independent correlation between hepatic FGF21 mRNA level and HCV RNA level. In an HCV replicon study, PPARα agonist inhibits HCV replication in Huh7/Rep-Feo cells.52 PPARα agonist reduces serum HCV RNA titers in patients.53, 54 Whether this effect is mediated by FGF21 warrants further study because FGF21 is a PPARα target.
The current study has a few limitations. First, the sample size is not large. A larger sample size would increase confidence in the findings. Second, expression was detected at the mRNA level because most livers were obtained from biopsy. However, most of the genes studied are regulated by nuclear receptors at the transcriptional level. Third, the number of genes studied was limited. We prioritized the assays by studying the expression of genes that have obvious roles in lipid, bile acid, and carbohydrate homeostasis. The cofactors studied are the most common ones for their receptive nuclear receptors. The ultimate approach would be microarray analysis using a large sample size. Fourth, direct comparison between normal liver and chronic hepatitis C patients with a drinking history was not done due to lack of matched sex. However, the changes caused by HCV and by alcohol may mask each other's effects. For example, PPARα mRNA was decreased in HCV-infected livers in comparison with normal livers. However, it was increased in HCV patients who had a history of alcohol drinking in comparison with those who were HCV infected but did not have a drinking history. Thus, it is likely that PPARα expression level is not different between the control and HCV/alcohol groups; however, the change to PPARα is significant and the interaction between HCV and alcohol drinking deserves attention.
The authors thank patients, physicians, nurses, Natali Navarro Cazarez, and Carly Thoma-Perry for their contribution to the KU Liver Center Tissue Bank. We also thank Zoe Raglow for her assistance in preparation of this manuscript.